“Colostethus” Ruthveni Group (Anura: Dendrobatidae) and Comments on the Distribution in the Sierra Nevada De Santa Marta, Colombia

Home , Atanquez, Colostethus, Colostethus inguinalis

SALAMANDRA 55(1) 27–38 15Acoustic February repertory 2019 ofISSN “Colostethus 0036–3375” ruthveni group

Acoustic repertory of the “Colostethus” ruthveni group (Anura: Dendrobatidae) and comments on the distribution in the Sierra Nevada de Santa Marta, Colombia

Juan David Jiménez-Bolaño1, Andrés Camilo Montes-Correa1, Freddy Polo-Córdoba1, Katherin Linares-Vargas1, Danilo Vergara-Gil2, Cesar L. Barrio-Amorós3 & Claudia Koch4

1) Grupo de Investigación en Manejo y Conservación de Fauna, Flora y Ecosistemas Estratégicos Neotropicales (MIKU), Universidad del Magdalena, Santa Marta, Colombia 2) Programa de Maestría en Ecología Acuática, Universidad de Zulia, Maracaibo, Venezuela 3) Doc Frog Expeditions, San José, Costa Rica 4) Zoologisches Forschungsmuseum Alexander Koenig, Bonn, Germany

Corresponding author: Claudia Koch, e-mail: [email protected]

Manuscript received: 3 July 2018 Accepted: 26 October 2018 by Stefan Lötters

Abstract. The Colostethus“ ” ruthveni group represents an endemic dendrobatid lineage in the Sierra Nevada de Santa Mar- ta (SNSM), an isolated massif in northeastern Colombia. This group has a complex taxonomic history and comprises the nominal species, from the northern and northwestern sector of the SNSM and an undescribed morph provisionally named as “Colostethus” sp. ruthveni-like, from the southeastern sector of this mountain massif. Herein, we describe the advertisement calls of members of the “C.” ruthveni group from several basins in the northern, northwestern, and southeastern sectors of the SNSM, and their geographic variation. In addition, we describe the courtship calls of the northwestern populations and provide comments on the distribution of the members of the group. The advertisement call of the C.“ ” ruthveni group consists of a trill with a long series of notes. All advertisement call traits show statistical differences among the assessed basins. The principal component analysis revealed overlap in call traits of individuals from the northwestern sector of the SNSM (“C.” ruthveni sensu stricto), while the populations of the southeastern sector (“C.” sp. ruthveni-like) and northern sector (“C.” cf. ruthveni) varied from these calls with respect to the dominant frequency and rate of notes per second, respectively. Acoustic divergence of advertisement calls of “C.” ruthveni sensu stricto and “C.” sp ruthveni-like is congruent with morphological and molecular evidence. Courtship call of “C.” ruthveni sensu stricto is a short partially fused pulsed squeak, followed by the usual advertisement call. This courtship call differs drastically in structure from the advertisement call, showing once again the complexity of the acoustic repertory of dendrobatoid frogs. Key words. Acoustic divergence, advertisement call, Colombian Caribbean, courtship call, geographic variation.

Introduction century, with the description of the back-riding tadpoles (Ruthven & Gaige 1915) and some aspects of the natural The Sierra Nevada de Santa Marta (SNSM) is an isolated history (Ruthven 1922) of “C.” ruthveni in the northern mountain massif in northeastern Colombia, in the de- and northwestern sectors of the SNSM, under the identity partments of Magdalena, Cesar and La Guajira. Edaphic of Hyloxalus subpunctatus (Cope, 1899). and climatic conditions of this mountain system promote The formal description ofColostethus “ ” ruthveni by speciation processes, resulting in an eco-geographic area Kaplan (1997) was based on specimens mainly collected with high endemicity (Carbonó & Lozano-Contreras in the northwestern sector of the SNSM (type locality: que- 1997, Lynch et al. 1997). According to Armesto & Señaris brada Viernes Santo, Santa Marta, Magdalena department) (2017), the SNSM has the highest proportion of endemic in the early 20th century and housed in the Museum of anurans within the northern Andes. One of these endemic Zoology of the Michigan University (UMMZ). Later, some lineages is the “Colostethus” ruthveni group, composed of contributions to the distribution (González-Maya et al. the homonym species (Kaplan, 1997) and at least one un- 2011, Granda-Rodríguez et al. 2014) and feeding ecology described morph (Lynch et al. 1997, Grant et al. 2017). (Blanco-Torres et al. 2014) of “C.” ruthveni were pub- The complex history of this group started in the early 20th lished. Rueda-Almonacid et al. (2008) provided a brief

© 2019 Deutsche Gesellschaft für Herpetologie und Terrarienkunde e.V. (DGHT), Mannheim, Germany Available at http://www.salamandra-journal.com 27 Juan David Jiménez-Bolaño et al. morphological description of an undescribed dendro from Los Besotes corresponds to an undescribed species, batoid species, temporarily named “Allobates sp.” from provisionally naming it “Colostethus” sp. ruthveni-like. a locality in the southeastern foothills of the SNSM (Los The importance of acoustic signals in intraspecific social Besotes Eco-park, Valledupar, Cesar department). In con- functions of anurans has been highlighted as being poten- trast, Anganoy-Criollo (2012, therein see Appendix A) tially useful for species recognition and taxonomic deci- and Granda-Rodríguez et al. (2014) referred to speci- sion making (Padial et al. 2009, Forti et al. 2017, Köhler mens from another locality in the southeastern sector of et al. 2017). In allopatric speciation scenarios, call differ- SNSM (Nabusímake, Valledupar, 22.7 km in straight line of ences are tricky to interpret as they do not necessarily in- Los Besotes) as “C.” ruthveni. This situation generated un- dicate specific distinctness, because separated populations certainty about the taxonomic status of the populations of of the same species can adapt to different environmental the southern region of the SNSM. conditions by modifying their vocalizations (Bernal et In a recent phylogenetic reconstruction of the Dendro al. 2005, Köhler et al. 2017). Nevertheless, we here start batoidea, Grant et al. (2017), proposed that the “Colo from the initial assumption that interspecific variation in stethus” ruthveni group is not related with the “coloste bioacoustic characters exceeds intraspecific variation (cf. thines” (or any other group of dendrobatoid frogs with Köhler et al. 2017). To elucidate the taxonomic status and cryptic coloration), but constitutes an undescribed genus the delimitation of species within the “C.” ruthveni group, closely related to toothless dendrobatines (tribe Dendro- this research was conducted in five hydrographic basins batini sensu Grant et al. 2017). The authors agreed with of the SNSM with the objectives (1) to describe the adver- Rueda-Almonacid et al. (2008) in that the population tisement call and the geographic variation of call traits of

Figure 1. Referred localities with presence of members of “Colostethus” ruthveni group in the SNSM (departments of Cesar, La Guajira and Magdalena, Colombia). Toribio basin (violet): 1. Viernes Santo creek (type locality), 2. La Tagua Village, 3. Hacienda El Aserrío, 4. Hacienda El Recuerdo; Gaira basin (green): 5. Arimaca, 6. El Berro creek, 7. Pozo Azul, 8. Hacienda La Victoria, 9. El Campano Village; Los Rodríguez basin (blue): 10. Las Tinajas Village; Ancho basin (red):11. Tugueca Village, 12. Wimangaga Reserve, 13. San Miguel Village; Cesar basin (yellowish): 14. Atanquez Village, 15. Los Besotes Eco-park, 16. Donachuí Village, 17. Nabusímake Village, 18. Pueblo Bello Village, 19. Don Diego Village.

28 Acoustic repertory of “Colostethus” ruthveni group

Table 1. Localities where calls of the members of the “Colostethus” ruthveni group were recorded. *Information on the courtship call in parenthesis.

Elevation N N Locality Geographic coordinates (Datum WGS84) Hour Temperature (°C) (m a.s.l.) males calls Gaira river basin El Campano Village 11°6’41.41”N, 74°5’36.24”W 1450 1 2 11:15–11:20 28.3 Hacienda La Victoria 11°7’44.10”N, 74°5’32.40”W 900 3 8 08:40–09:00 24.7 Pozo Azul 11°7’58.72”N, 74°6’07.58”W 836 2 12 09:30–09:50 25.1 Arimaca* 11°7’33.56”N, 74°7’08.93”W 723 3 (1) 12 (4) 15:15–15:30 24.6 El Berro creek 11°8’18.33”N, 74°7’00.88”W 620 2 12 14:45–15:00 26.5 Toribio river basin La Tagua Village 11°4’55.00”N, 74°2’22.70”W 1500 3 11 08:30–08:40 23.9 Hacienda El Aserrío 11°4’56.80”N, 74°6’41.90”W 1400 4 10 15:40–15:50 26.5 Hacienda El Recuerdo 11°5’14.00”N, 74°2’53.00”W 1200 1 5 09:20–09:30 25.6 Los Rodríguez creek basin Las Tinajas Village 11°17’03,80’’N, 74°4’00,20”W 735 5 11 10:15–10:30 28.2 Ancho river basin Wimangaga Reserve 11°8’18.42”N, 73°32’45.94”W 700 2 11 09:00–09:15 25.6 Tugueca Village 11°8’49.11”N, 73°30’32.68”W 450 1 12 10:20–10:30 29.3 Cesar river Basin Donachuí Village 10°42’47.60”N, 73°27’20.64”W 1436 3 6 16:40–16:50 27.2 the members of the “C.” ruthveni group, (2) to evaluate the Call collection and analyses potential acoustic divergence between “C.” ruthveni from the northern and northwestern sector of the SNSM and Calls of 30 males with a duration of in total 40 min. were “C.” sp. ruthveni-like, (3) to provide a description of the recorded with a Tascam DR05 recorder during seven field courtship call from populations of the northwestern sector trips between December 2014 and April 2017; recordings of SNSM, and (4) to add information on the distribution of were made at 08:00–12:00 h and 15:00–17:00 h, at an ap- the group members. proximate distance of 50 cm from the calling male (Ta- ble 1). Air temperature was measured with a digital ther- mo-hygrometer Extech RH210. Voucher specimens were Materials and methods not collected and each male at each site was only recorded Study area once to avoid pseudoreplication. To complement the taxonomic framework of the calls, available specimens of the Advertisement calls (AC) of specimens of the “Colostethus” “Colostethus” ruthveni group in the amphibian collection ruthveni group were recorded at different localities in the of the Centro de Colecciones Biológicas de la Universidad northern, northwestern, and southeastern sectors of the del Magdalena (CBUMAG:ANF) were revised, including SNSM (Table 1, Fig. 1). In the northwestern sector (Santa specimens from the same (or nearby) localities of the re- Marta district, Magdalena department), recordings were corded calls. conducted at five localities in the middle Gaira river basin, All calls were analyzed with Praat (v. 6.0.13) for win- at three localities in the middle Toribio river basin, and at dows (Boersma, P. & D. Weenink 2007: Praat: Doing pho- one locality in the upper Los Rodríguez creek basin. In the netics by computer; available from http://www.fon.hum. northern sector (Dibulla municipality, La Guajira depart- uva.nl/praat/; accessed April 2015) at a sampling frequency ment), calls were recorded at two localities from the lower of 44.1 kHz and 16 bits resolution. We consider continuous and middle Ancho river basin. In the southeastern sector vocalizations performed by males to attract females as AC, (Valledupar municipality, Cesar department), recordings and particular vocalizations emitted by males in the pres- were conducted at one locality in the upper Cesar river ence of females as CC (Wells 2007, Köhler et al. 2017). basin. The courtship call (CC) was recorded in Arimaca, Each call was extracted from the original recording as an in the middle Gaira river basin in the northwestern sector individual file. For each call, the following traits were meas- (Table 1, Fig. 1). The midlands of the northwestern sector ured: call duration (in seconds = s), number of notes per and the lowlands of the northern sector of the SNSM have call, rate of notes per second (notes/s), note duration (s), humid forest formations, while the southwestern sector internote interval (s), dominant frequency (Hz), and vis- shows vegetation related to tropical dry forest formations ible harmonics. Means, standard deviation (SD), and coef- (Hernández-Camacho et al. 1992). ficient of variation (CV) were calculated for each call trait.

29 Juan David Jiménez-Bolaño et al.

Call traits with CV ≤ 0.05 were considered as static traits ate data analyses. R package version 1.0.5; available from (cf. Gerhardt 1991). We follow the terminology proposed https://CRAN.R-project.org/package=factoextra; accessed and revised by Lötters et al. (2003) for advertisement call June 2017). Spectrograms and oscillograms were obtained types and Köhler et al. (2017) for call traits. with the ‘Seewave’ package for Cran R (Sueur et al. 2008). A Kruskal-Wallis test (H) was performed to evaluate the potential geographic variation of the AC of the members of the “C.” ruthveni group, in order to refuse the null hypoth- Results esis that no differences in call traits (note duration, inter- Advertisement call note interval, rate of notes per second, and dominant frequency) exist between different basins. As post hoc anal- In the first instance, no obvious qualitative differences ysis, a pairwise Wilcoxon test was conducted. Both tests were detected in the AC of the populations assigned to the were performed with the ‘basic’ package for the Cran R “Colostethus” ruthveni group (Fig. 2). Nevertheless, sig- platform, analyzing each note individually. To reduce the nificant variation was detected in certain quantitative call variation of call traits and visualize potential cluster in traits among individuals in the basins assessed (Table 2). terms of similarity, a principal component analyses (PCA) The Kruskal-Wallis test showed significant differences in of the mean values of call traits was performed for each note duration between the different populations (H = 1313, male. This analysis was conducted with the ‘Factoextra’ df [degrees of freedom] = 4, p < 0.001, Fig. 3a). According package for Cran R (Kassambara, A. & F. Mundt 2017: to the grouping by significance of the pairwise Wilcoxon Factoextra: extract and visualize the results of multivari- test, all basins showed differences among themselves (all

Figure 2. Spectrograms and oscillograms of the advertisement calls of “Colostethus” ruthveni sensu stricto (A, El Berro creek, Minca Village, Gaira basin, Santa Marta, Magdalena), “Colostethus” cf. ruthveni (B, Tugueca village, Río Ancho basin, Dibulla, La Guajira), and “Colostethus” sp. ruthveni-like (C, Donachuí village, Cesar basin, Valledupar, Cesar). Two seconds fragments (left) and entire calls (right) for each group. Grayish area shows the two seconds fragment within the call.

30 Acoustic repertory of “Colostethus” ruthveni group

Table 2. Variation of the advertisement call traits of the members of the “Colostethus” ruthveni group by basins.

Basin N Mean SD Minimums Maximums CV Grouping by Wilcoxon pairwise test Note duration (s) Ancho 478 0.029 0.005 0.017 0.083 0.183 a Gaira 1203 0.067 0.024 0.02 0.17 0.35 b Cesar 119 0.025 0.01 0.009 0.046 0.397 c Los Rodríguez 312 0.059 0.011 0.028 0.098 0.181 d Toribio 482 0.054 0.014 0.02 0.097 0.255 e Kruskal-Wallis test H = 1313, df = 4, p < 0.001 Internote interval (s) Ancho 455 0.090 0.016 0.056 0.189 0.179 a Gaira 1159 0.122 0.039 0.053 0.445 0.322 b Cesar 113 0.153 0.043 0.074 0.419 0.278 c Los Rodríguez 302 0.144 0.027 0.045 0.247 0.188 c Toribio 457 0.203 0.081 0.064 0.625 0.399 d Kruskal-Wallis test H = 1265.1, df = 4, p < 0.001 Rate of notes per second (notes/s) Ancho 23 8.673 0.760 7.607 9.713 0.088 a Gaira 45 6.003 1.629 4.172 12.953 0.271 b Cesar 6 5.603 1.069 3.701 6.564 0.191 bc Los Rodríguez 11 4.998 0.101 4.847 5.142 0.02 d Toribio 26 4.462 1.487 2.643 9.631 0.333 c Kruskal-Wallis test H = 64.263, df = 4, p < 0.001 Dominant frequency (Hz) Ancho 478 4909.584 279.119 4180 5307 0.057 a Gaira 1203 5030.381 195.643 4600 5539 0.039 b Cesar 119 4301.160 90.273 3979 4487 0.021 c Los Rodríguez 312 5742.763 95.075 5402 5944 0.017 d Toribio 482 4732.770 279.575 4157 5238 0.059 e Kruskal-Wallis test H = 1302.2, df = 4, p < 0.001 pairwise tests with p < 0.001). Note duration was a dynam- Table 3. Results of the principal components analysis, summariz- ic call trait, showing a considerable variation even among ing the variation of the advertisement calls of the members of the individuals within each basin (CV = 0.181–0.397). The re- “Colostethus” ruthveni group. sult for internote intervals was similarly variable between most basins (H = 1265.1, df = 4, p < 0.001, Fig. 3b), with Call trait Dim.1 Dim.2 Dim.3 only one pair of basins that did not differ from each oth- Note duration (s) -0.175 0.874 -0.445 er (Cesar vs. Los Rodríguez, p = 0.08). This call trait also Internote interval (s) -0.947 -0.143 0.194 exhibited considerable variation among individuals within each basin (CV =0.179–0.399). Rate of notes per second (notes/s) 0.449 0.737 0.503 The differences in the rate of notes per second were re- Dominant frequency (Hz) 0.903 -0.347 -0.133 markable between the basins (H = 64.263, df = 4, p < 0.001, Eigenvalues 1.945 1.448 0.507 Fig. 3c), but presented a slightly smaller variation than Explained variance (%) 48.633 36.206 12.676 the call traits mentioned above within each basin (CV = 0.088–0.333). There were no differences between Gaira summarized by three dimensions of the PCA (Table 3, vs. Cesar (p = 0.804) and between Cesar vs. Toribio (p = Fig. 4). Internote interval and rate of notes per second were 0.057). Finally, the dominant frequency also differed be- mainly explained by dimension 1, while note duration and tween all basins (H = 1302.2, df = 4, p < 0.001, Fig. 3d), al- dominant frequency were mainly explained by dimensions though this call trait showed the smallest variation (almost 2 and 3. The first two dimensions explained 84.84% of the static) within each basin (CV < 0.06 in all basins). variance in the PCA (Table 3). Results of the PCA reflect Variation of note duration, internote interval, rate of that call traits of individuals from Gaira and Los Rodríguez notes per second, and dominant frequency of the AC of basins are heavily overlapping, whereas individuals from the members of the “Colostethus” ruthveni group was well Gaira and Toribio basins showed only a slight overlap, and

31 Juan David Jiménez-Bolaño et al. those from Cesar and Ancho basins were positioned sep- The AC of members of the Colostethus“ ” ruthveni group arately, varying in their dominant frequency and rate of consists of a long series of trills with pulsating notes and notes per second, respectively (Fig. 4). These results sup- modulated amplitude (Table 4, Fig. 2). In “C.” ruthveni sen- port the divergence of populations from the northwest- su stricto the mean duration of the AC was 4.868 s (SD = ern sector of the SNSM (hereinafter named “C.” ruthveni 2.358 s), with a mean of 25 notes per call (SD = 11 notes), sensu stricto, Fig. 5a) and the southwestern sector (“C.” sp. a mean note duration of 0.063 s (SD = 0.021 s), a mean ruthveni-like, Fig. 5b). Surprisingly, the PCA revealed par- rate of 5.380 notes per second (SD = 1.622 notes/s), and in- ticular acoustic features of males from the northern sector ternote intervals of 0.145 s (SD = 0.061 s). Call frequency (hereinafter named C.“ ” cf. ruthveni, Fig. 5c). was in the spectral band between 3,000–16,000 Hz. Notes

Figure 3. Geographic variation of note duration (A), internote interval (B), rate of notes per second (C), and dominant frequency (D). Colors and names of X axis correspond to information in Table 1 and Figure 1.

Figure 4. Variation of the advertisement call traits explained by two principal components. Colors and names correspond to information in Table 1 and Figure 1.

32 Acoustic repertory of “Colostethus” ruthveni group

Table 4. General data of the advertisement calls of the members of the “Colostethus” ruthveni group.

Call trait N Mean SD Minimums Maximums CV “Colostethus” ruthveni sensu stricto Call duration (s) 82 4.868 2.358 1.036 12.464 0.484 Notes per call 82 25 11 5 52 0.452 Rate of notes per second (notes/s) 82 5.380 1.622 2.643 12.953 0.302 Note duration (s) 1997 0.063 0.021 0.020 0.170 0.332 Internote interval (s) 1918 0.145 0.061 0.045 0.625 0.421 Dominant frequency (Hz) 1997 5069.847 377.427 4157 5944 0.074 Harmonic 1 1857 7648.673 581.536 5448 9947 0.076 Harmonic 2 850 9594.842 611.093 4988 12949 0.064 Harmonic 3 504 12680.165 456.792 11045 15142 0.036 Harmonic 4 493 14970.233 354.457 13745 17139 0.024 “Colostethus” sp. ruthveni-like Call duration (s) 6 3.363 1.578 1.351 4.773 0.469 Notes per call 10 20 11 5 30 0.546 Rate of notes per second (notes/s) 10 5.603 1.069 3.701 6.564 0.191 Note duration (s) 119 0.025 0.010 0.009 0.046 0.397 Internote interval (s) 113 0.153 0.043 0.074 0.419 0.278 Dominant frequency (Hz) 119 4301.160 90.273 3979 4487 0.021 Harmonic 1 119 8196.504 104.165 7925 8404 0.013 Harmonic 2 119 12091.546 109.780 11794 12314 0.009 Harmonic 3 119 15982.378 110.625 15688 16187 0.007 Harmonic 4 119 19879.647 110.643 19590 20112 0.006 “Colostethus” cf. ruthveni Call duration (s) 23 2.391 0.340 1.848 3.086 0.142 Notes per call 23 21 4 15 29 0.182 Rate of notes per second (notes/s) 23 8.673 0.760 7.607 9.713 0.088 Note duration (s) 478 0.029 0.005 0.017 0.083 0.183 Internote interval (s) 455 0.090 0.016 0.056 0.189 0.179 Dominant frequency (Hz) 478 4909.584 279.119 4180 5307 0.057 show ascending frequency modulation, with a mean domi- Courtship call nant frequency of 5,069.847 Hz (SD = 377.427 Hz). Notes have one to four visible harmonics at 7,648.673 Hz (SD = The description of the CC of “Colostethus” ruthveni sensu 581.536 Hz), 9,594.842 Hz (SD = 611.093 Hz), 12,680.165 Hz stricto was based on four calls of one male. This vocaliza- (SD = 456.792), and 14,970.233 Hz (SD = 345.457 Hz). When tion consists of a short squeak with partially fused pulses evaluating the variation of each call trait of “C.” ruthveni and modulated amplitude (Fig. 6). The mean duration of sensu stricto, it is evident that the dominant frequency was the CC is 0.124 s (SD = 0.038 s, range = 0.077–0.293 s) with the only almost static feature (CV = 0.074), while the oth- a frequency between 4,415–12,300 Hz and a mean domi- ers showed a considerable variation (CV > 0.3). nant frequency of 4,706.6 Hz (SD = 252.204 Hz, range = The AC of Colostethus“ ” sp. ruthveni-like differs consid- 4,489–5,093 Hz). Harmonics are not visible. The CC is fol- erably from the AC of “Colostethus” ruthveni sensu stric- lowed by a silence interval with a mean duration of 0.566 s to by its lower dominant frequency (mean = 4,301.160 Hz, (SD = 0.525, range = 4,489–5,093 Hz) before the AC starts. SD = 90.273 Hz) and shorter notes (mean = 0.025, SD = 0.010 s). All analyzed notes show four visible harmonics at 8,196.504 Hz (SD = 104.165 Hz), 12,091.546 Hz (SD = 109.780 Discussion Hz), 15,982.378 Hz (SD = 110.625 Hz), and 19,879.647 Hz (SD Advertisement call = 110.643 Hz). The AC of C.“ ” cf. ruthveni differs from those previously mentioned by its higher rate of notes per second All call traits evaluated for the members of the “Colostethus” (mean = 8.673 notes/s, SD = 0.760 notes/s) and its shorter ruthveni group differed among the basins. Although the internote intervals (mean = 0.090 s, SD = 0.016 s). None of effect of ambient temperature and male size could not be the ACs of “C.” cf. ruthveni showed visible harmonics. evaluated, it can be assumed that all the variation produced

33 Juan David Jiménez-Bolaño et al. by these external variables is summarized in the CVs cal- thesis that northwestern and southeastern populations of culated for each call trait in each basin. The variation in the “C.” ruthveni group are not conspecific is reinforced. the ACs of allopatric populations may be due to potentially Advertisement calls of the members of the “C.” ruth different environmental conditions Köhler( et al. 2017). veni group consist of a relatively long sequence of trills. For example, clinal variations have been described in the However, this AC type is not present in the tribe Dendro- ACs of other dendrobatoids, such as Rheobates palma batini, the clade most closely related to the “C.” ruthveni tus (Werner, 1899) from the Eastern Andes of Colom- group (Grant et al. 2017). In fact, the AC of members of bia (Bernal et al. 2005) and Oophaga pumilio (Schmidt, the “C.” ruthveni group is more similar in structure (and 1857) from Central America (Pröhl et al. 2007). On the to the human ear) to the AC of Phyllobates Bibron, 1840, other hand, early intrapopulation divergence has been evi- aromobatids, colostethines, and hyloxalines. This supports denced in O. granulifera (Taylor, 1958) from Costa Rica, that the trills have appeared several times in the evolution- where two lineages show acoustic differentiation (Brusa et ary history of Dendrobatoidea (Lötters et al. 2003, Erdt- al. 2013). The acoustic differentiation ofC. “ ” ruthveni sen- mann & Amézquita 2009, Santos et al. 2014). su stricto and “C.” sp. ruthveni-like is congruent with the Comparisons of the ACs of the “C.” ruthveni group are morphological and genetic divergence (Rueda-Almona limited by the few calls described for geographically close cid et al. 2008, Grant et al. 2017). Therefore, the hypo or morphologically similar species. In structure, the ACs of

Figure 5. Specimens of the “Colostethus” ruthveni group. “Colostethus” ruthveni sensu stricto from Hacienda la Victoria, Santa Marta, Magdalena department (A). “Colostethus” sp. ruthveni-like from Donachuí Village, Valledupar, Cesar department (B). “Colostethus” cf. ruthveni from Wimangaga reserve, Dibulla, La Guajira department (C, specimen CBUMAG:ANF:00939). Small creek; typical habitat of members of the “Colostethus” ruthveni group (D).

34 Acoustic repertory of “Colostethus” ruthveni group these frogs are very similar to those of other trans-Andean braccata (Steindachner, 1864), Allobates femoralis (Bou- dendrobatoids such as Allobates talamancae (Cope, 1875 lenger, 1884 “1883”) and A. talamancae. However, CCs “1876”), A. ignotus Anganoy-Criollo, 2012, R. palmatus, of these species differ from those of C. “ ” ruthveni sensu and Colostethus panamensis (Dunn, 1933). However, the stricto by their duration and dominant frequency (P. bi dominant frequencies of A. talamancae (4,865–5,354 Hz, color: 0.6 s, 2,375 Hz, Zimmerman & Zimmerman 1985; Lechelt et al. 2014) and A. ignotus (4,562–5,312 Hz, Gran- A. braccata: 0.043 s, 3,609.8–3,778.8 Hz, Forti et al. 2010; da-Rodríguez et al. 2018) exceed the dominant frequen- A. femoralis: without specific values, see spectrograms of cies of “C.” sp. ruthveni-like (3,979–4,487 Hz) and “C.” cf. Weygoldt 1980; A. talamancae: 0.2 s, 1,500–4,200 Hz, ruthveni (4,180–5,307 Hz), while they are slightly lower Kollarits et al. 2017). than the dominant frequency of “C.” ruthveni sensu stric- Unlike in “C” ruthveni sensu stricto, the CCs of some to (4,177–5,944 Hz). In contrast, the dominant frequency dendrobatoids are not very different from their ACs. In of R. palmatus (2,700 Hz, Lüddecke 1999) is consider- some cases, the CC consists of a continuous sequence of ably lower than that of any member of the “C.” ruthveni trills, but with notes of a different shape from the typical group. The dominant frequency of C. panamensis (4,126– AC, such as in R. palmatus (20 long notes, 2,000–2,800 Hz, 4,547 Hz, Wells 1980a) is similar to “C.” sp. ruthveni- Lüddecke 1999). Other species have a CC with simi- like, but the frequency range is smaller than in “C.” ruth lar notes to those of the AC, differing only in the number veni sensu stricto and “C.” cf. ruthveni. For other dendro of notes per call and the rate of repetition. For example, batoids with trans-Andean distribution, such as Allo Mannophryne trinitatis (Garman, 1888 “1887”), Anomalo bates niputidea Grant, Acosta-Galvis & Rada, 2007, glossus stepheni (Martins, 1989) and Allobates marchesia A. wayuu (Acosta-Galvis, Coloma & Cuentas, 1999), nus (Melin, 1941) produce ACs composed of few notes, Aromobates aff. saltuensis (Rivero, 1980 “1978”), A. toku while the CCs consists of a longer sequence of similar ko Rojas-Runjaic, Infante-Rivero & Barrio-Amorós, notes, with lower (A. stepheni: AC = 7–8 notes per call, 2011, Colostethus inguinalis (Cope, 1868) and Dendrobates 4,600–4,800 Hz vs CC = 34 notes per call, 3,500–4,500 Hz, truncatus (Cope, 1861 “1860”) (Acosta-Galvis et al. 1999, Juncá 1998) or equal (M. trinitatis: 2 notes per call vs 3–27 Grant 2004, Grant et al. 2007, Rojas-Runjaic et al. 2011, notes per call, 4,300–4,600 Hz, Wells 1980b, La Marca Anganoy-Criollo 2012, Gualdrón-Duarte et al. 2016), 1994; A. marchesianus: 1 note per call vs 30 notes per call, advertisement call descriptions of have not been published. 4,500–7,000 Hz, Juncá 1998) dominant frequencies. In general, anurans have a lower dominant frequency in the CC as compared to the AC (Toledo et al. 2015). How- Courtship call ever, there are particular cases such as Anomaloglossus bee bei (Noble, 1923), where the CC and the AC have the same The courtship call of Colostethus“ ” ruthveni sensu stricto is structure (note triplets), but the dominant frequency of the a squeaking noise, emitted only during the close encounter AC is higher than that of the CC (4,630–5,730 Hz vs 3,750– of males and females. This type of call differs considerably 4,590 Hz, Pettit et al. 2012). On the other hand, some in structure from the AC in that it is much shorter but with species such as Hyloxalus toachi (Coloma, 1995) present longer partially fused notes, and a slightly lower dominant two types of CCs, which have dominant frequencies in the frequency. These characteristics have also been described same range (4,000–4,300 Hz) and the same note structure for the CC of Phyllobates bicolor Bibron 1840, Ameerega as in the AC, but differ in the number of notes per call and

Figure 6. Courtship call of “Colostethus” ruthveni sensu stricto from Arimaca, Marinca Creek, Minca Village, Gaira basin. Half-second fragment (left) and full recording including subsequent advertisement call (right). Grayish area shows the half-second fragment within the full recording.

35 Juan David Jiménez-Bolaño et al. the rate of repetition (AC = 10–51 notes per call, two calls were collected (Ruthven 1922, Kaplan 1997). Although per minute; CC1 = 32–72 notes per call, one call per minute; Rivero (1963, foot note, p. 119) mentioned that San Miguel CC2 = 1–2 notes per call, 1–6 calls per minute, Quiguan- and Don Diego are not the same place, several subsequent go-Ubillús & Coloma 2008). Although the CC is poor- publications referred to these sites as being the same or ly known in anurans in general, the scarce literature avail- very close localities situated in the Magdalena department able shows that dendrobatoids present highly diversified (Ruiz-Carranza et al. 1994, Kaplan 1997, Granda-Ro- CCs, probably related to the complexity of social behavior dríguez et al. 2014). However, Don Diego is a coastal town described for these frogs (Dole & Durant 1974, Wells (at sea level) located at the mouth of the homonymous ba- 1980a, b, Wells 2007). sin, northern sector of the SNSM, Santa Marta, Magdalena department. No members of the “C.” ruthveni group have so far been reported from this site, although a distribution Taxonomic remarks in the middle basin of the Don Diego River is very likely. On the other hand, San Miguel is an indigenous town lo- To restrict the populations from the northwestern sec- cated in mid-altitudes (1,600 m a.s.l.) in the northern sector of the SNSM as “Colostethus” ruthveni sensu stricto, tor of the SNSM, on the Macotama Creek, a tributary of the following aspects were taken into account: (1) Viernes the Ancho river in its upper basin, Dibulla, nowadays in Santo creek is a tributary of the Toribio river, (2) most of La Guajira department. Although we could not review the the type series (14 of 16 specimens) described by Kaplan type series of the species, the paratypes from Ancho basin (1997) come from the Gaira and Toribio basins, (3) Gaira, probably also correspond to “C.” cf. ruthveni. Toribio and Los Rodríguez basins are the geographically closest basins, (4) there are no obvious morphological differences (sensu Kaplan 1997) between specimens coming Acknowledgments from these basins. We are grateful to the people from Minca village, the people from Unfortunately, we were unable to obtain specimens the hacienda La Alicia, Sebastian Ballesteros at El Faunal Re- from Donachuí (upper Cesar basin). Nevertheless, males serve, Carlos Fernández at Wimangaga Reserve and the Zala- recorded in this locality are morphologically very simi- bata family for allowing us to collect field data in their homes. We lar to specimens from Pueblo Bello (33.9 km southwest of thank Hernán Granda-Rodríguez, Liliana Saboyá-Acosta, Donachuí) and Atanquez (11.7 km east of Donachui), two Yuli Castro, Efraín Rada, Miguel Gutiérrez, David Arau- villages of Valledupar (Cesar department). Males from jo and Jeison Vanegas for their assistance during the field work. Donachuí, Pueblo Bello, and Atanquez coincide in an ex- We further thank Kenedith Mendez for advising us on statis- ceptional character of sexual dimorphism, with a darker tical issues. Andrea Cotes and Ruben Fontalvo kindly pro- coloration of the gular region (absent in males of “C.” ruth vided comments on the manuscript. Finally, we are indebted to Roberto Guerrero and Amanda Berbén, from the Centro de veni sensu stricto). This character was described by Rue- Colecciones Biológicas de la Universidad del Magdalena for their da-Almonacid et al. (2008) for “C.” sp. ruthveni-like (as unconditional support. “Allobates sp.”) from Los Besotes. Results of the PCA agree with the hypothesis on the divergence between “C.” ruth veni sensu stricto and “C.” sp. ruthveni-like proposed by References Grant et al. (2017). Certainly, the specimens of “C.” ruth veni reported from Nabusímake (Anganoy-Criollo Acosta-Galvis, A. R., D. A. Cuentas & L. A. Coloma (1999): 2012, Granda-Rodríguez et al. 2014) also correspond to Una nueva especie de Colostethus (Anura: Dendrobatidae) de la región del Caribe de Colombia. – Revista Colombiana de la “C.” sp. ruthveni-like. Academia de Ciencias Exactas, Físicas y Naturales, 23: 225– Populations of the “Colostethus” ruthveni group from 230. the northern sector of SNSM are still poorly known. We Anganoy-Criollo, M. (2012): A new species of Allobates (An consider the specimens from Ancho river basin as “C.” cf. ura, Dendrobatidae) from the western flank of the Serranía de ruthveni by their acoustic differences, and the considerable Perijá, Colombia. – Zootaxa, 3308: 49−62. geographic distance to “C.” ruthveni sensu stricto. Further- Armesto, L. O. & J. C. Señaris (2017): Anuros del norte de los more, for two specimens from Wimangaga reserve, mid- Andes: patrones de riqueza de especies y estado de conserva- dle Ancho river basin, Dibulla, La Guajira department, the ción. − Papéis Avulsos de Zoologia, 57: 491−526. disc of toe V does not reach the anterior border of the mid- Bibron, G. (1840): Historia Fisica, Politica y Natural de la Isla de dle subarticular tubercle of the toe IV when toes are ad- Cuba, Segunda Parte. Historia Natural. Tomo VIII. Atlas de pressed (at the same level in “C.” ruthveni sensu stricto). Zoología. − Arthus Bertrand, Paris. Nevertheless, additional specimens from the basin are re- Bernal, X. E., C. Guarnizo & H. Lüddecke (2005): Geographic quired, to determine if this condition is a consequence of variation in advertisement call and genetic structure of Colo the state of preservation of the specimens or if it represents stethus palmatus (Anura, Dendrobatidae) from the Colombi- a morphological character established in this population. an Andes. – Herpetologica, 41: 395–408. A detail that requires clarification is the geographic local- Blanco-Torres, A., M. P. Ordóñez-Pachón & M. C. Franco- ization of the “Mountains west [or east] of San Miguel and Rozo (2014): Colostethus ruthveni (Santa Marta poison arrow Don Diego”, a locality where two paratypes of “C.” ruthveni frog). Diet. − Herpetological Review, 45: 476.

36 Acoustic repertory of “Colostethus” ruthveni group

Boulenger, G. A. (1884 “1883”): On a collection of frogs from drobatid frog Colostethus ruthveni Kaplan, 1997 (Anura: Den- Yurimaguas, Huallaga River, Northern Peru. – Proceedings of drobatidae) in the Sierra Nevada de Santa Marta, Colombia. the Zoological Society of London, 1883: 635–638. – Check List, 10: 674–676. Brusa, O., A. Bellati, I. Meuche, N. I. Mundy & H. Pröhl Granda-Rodríguez, H. D., A. C. Montes-Correa, J. D. Jimén- (2013): Divergent evolution in the polymorphic granular poi- ez-Bolaño, M. Anganoy-Criollo. (2018): Natural history son-dart frog, Oophaga granulifera: genetics, coloration, ad- and conservation of the nurse frog of the Serranía de Perijá vertisement calls and morphology. – Journal of Biogeography, Allobates ignotus (Anura: Aromobatidae) in northeaster Co- 40: 394–408. lombia. – Acta Herpetologica, 13: 51–64. Carbonó, E. & G. Lozano-Contreras (1997): Endemismos y Grant, T. (2004): On the Identities of Colostethus inguinalis otras singularidades de la Sierra Nevada de Santa Marta; po- (Cope, 1868) and C. panamensis (Dunn, 1933), with Com- sibles causas de origen y necesidad de conservarlos. – Revista ments on C. latinasus (Cope, 1863) (Anura: Dendrobatidae). de la Academia Colombiana de Ciencias Exactas, Físicas y Na- – American Museum Novitates, 3444: 1–24. turales, 21: 409−419. Grant, T., A. R. Acosta-Galvis, M. Rada (2007): A name for Coloma, L. A. (1995): Ecuadorian frogs of the genus Colostethus the species of Allobates (Anura: Dendrobatoidea: Aromobati- (Anura: Dendrobatidae). – Miscellaneous Publication. – Mu- dae) from the Magdalena Valley of Colombia. – Copeia, 2007: seum of Natural History, University of Kansas, 87: 1–7. 844–854. Cope, E. D. (1861 “1860”): Descriptions of reptiles from tropical Grant, T., M. Rada, M. Anganoy-Criollo, A. Batista, P. H. America and Asia. – Proceedings of the Academy of Natural Dias, A. M. Jeckel, D. J. Machado & J. V. Rueda-Almona Sciences of Philadelphia, 12: 368–374. cid (2017): Phylogenetic systematics of dart-poison frogs and Cope, E. D. (1868): An examination of the Reptilia and Batrachia their relatives revisited (Anura: Dendrobatoidea). – South obtained by the Orton Expedition to Equador and the Upper American Journal of Herpetology, 12: S1−S90. Amazon, with notes on other species. – Proceedings of the Gualdrón-Duarte, J., V. F. Luna-Mora, M. Rivera-Correa & Academy of Natural Sciences of Philadelphia, 20: 96–140. T. Kahn (2016): Yellow-striped Poison Frog Dendrobates trun Cope, E. D. (1875 “1876”): On the Batrachia and Reptilia of Costa catus (Cope, 1861 “1860”). – pp. 323–327 in: Kahn, T. R., E. La Rica. – Journal of the Academy of Natural Sciences of Phila- Marca, S. Lötters, J. Brown, E. Twomey & A. Amézquita delphia Series 2, 8: 93–154. (eds): Aposematic poison frogs (Dendrobatidae) of the An- dean countries: Bolivia, Colombia, Ecuador, Perú and Vene- Cope, E. D. (1899): Contributions to the herpetology of New Gra- zuela. – Conservation International, Arlington. nada and Argentina, with descriptions of new forms, by Ed- ward D. Cope; a posthumous paper edited by J. Percy Moore. Hernández-Camacho, J. I., A. Hurtado-Guerra, R. Ortiz- – Philadelphia Museums, Science Bulletin, 1: 1–22. Quijano & T. Walschburger (1992): Origen y distribu- ción de la biota suramericana y colombiana. – pp. 55−104 in: Dole, J. W. & P. Durant (1974): Courtship behavior in Colo Halffter, G. I. (ed.): La diversidad biológica de Iberoaméri- stethus collaris (Dendrobatidae). – Copeia, 1974: 988–990. ca. – CYTED-D, Acta Zoológica Mexicana, México D.F. Dunn, E. R. (1933): Amphibians and reptiles from El Valle de An- Juncá, F. A. (1998): Reproductive biology of Colostethus stepheni ton, Panama. – Occasional Papers of the Boston Society of and Colostethus marchesianus (Dendrobatidae), with the de- Natural History, 8: 65–79. scription of a new anuran mating behavior. – Herpetologica, Erdtmann, L. & A. Amézquita (2009): Differential evolution of 54: 377–387. advertisement call traits in dart-poison frogs (Anura: Dendro- Kaplan, M. (1997): A new species of Colostethus from the Sierra batidae). – Ethology, 115: 801–811. Nevada de Santa Marta (Colombia) with comments on inter- Forti, L. R., C. Strüssmann & T. Mott (2010): Acoustic com- generic relationships within the Dendrobatidae. – Journal of munication and vocalization microhabitat in Ameerega brac Herpetology, 31: 369–375. cata (Steindachner, 1864) (Anura, Dendrobatidae) from Köhler, J., M. Jansen, A Rodríguez, P. J. R. Kok, L. F. Toledo, Midwestern Brazil. – Brazilian Journal of Biology, 70: 211–216. M. Emmrich, F. Glaw, C. F. B. Haddad, M. O. Rödel & M. Forti, L. R., T. R. Á. da Silva & L. F. Toledo (2017): The acous- Vences (2017): The use of bioacoustics in anuran taxonomy: tic repertoire of the Atlantic Forest Rocket Frog and its conse- theory, terminology, methods and recommendations for best quences for taxonomy and conservation (Allobates, Aromo- practice. – Zootaxa, 4251: 1–124. batidae). – ZooKeys, 692: 141−153. Kollarits, D., C. Wappl & M. Ringler (2017): The role of tem- Garman, S. (1888 “1887”): West Indian Batrachia in the Museum poral call structure in species recognition of male Allobates of Comparative Zoology. – Bulletin of the Essex Institute. Sa- talamancae (Cope, 1875) (Anura: Dendrobatidae). – Herpeto- lem, Massachusetts, 19: 13–16. zoa, 29: 115–124. Gerhardt, H. C. (1991): Female mate choice in treefrogs: static La Marca, E. (1994): Taxonomy of the frogs of the genus Manno and dynamic acoustic criteria. – Animal Behaviour, 42: 615–635. phryne (Amphibia: Anura: Dendrobatidae). – Publicaciones González-Maya, J. F, M. González, D. Zárate-Charry, F. de la Asociación de Amigos de Doñana, 4: 1–75. Charry, A. A. Cepeda, & S. A. Balaguera-Reina (2011): A Lechelt, S., W. Hödl & M. Ringler (2014): The role of spectral new population record and conservation assessment of the advertisement call properties in species recognition of male Santa Marta Poison Arrow Frog Colostethus ruthveni Kaplan, Allobates talamancae (Cope, 1875) (Amphibia: Aromobati- 1997 (Anura: Dendrobatidae) from Sierra Nevada de Santa dae). – Herpetozoa, 26: 139–150. Marta, Colombia. – Journal of Threatened Taxa,3 : 1633−1636. Lötters, S., S. Reichle & K. H. Jungfer (2003): Advertisement Granda-Rodríguez, H. D., L. P. Saboyá-Acosta, A. del Por- calls of Neotropical poison frogs (Amphibia: Dendrobatidae) tillo-Mozo & J. M. Renjifo (2014): Range extension of den- of the genera Colostethus, Dendrobates and Epipedobates, with

37 Juan David Jiménez-Bolaño et al.

notes on dendrobatid call classification. – Journal of Natural Ruthven, A. G. & G. T. Gaige (1915): The breeding habits of History, 37: 1899–1911. Prostheraspis subpunctatus Cope. – Occasional Papers of the Lüddecke, H. (1999): Behavioral aspects of the reproductive bi- Museum of Zoology, University of Michigan, 10: 1−7. ology of the Andean frog Colostethus palmatus (Amphibia: Santos, J. C., M. Baquero, C. L. Barrio-Amorós, L. A. Co- Dendrobatidae). – Revista de la Academia Colombiana de loma, L. K. Erdtmann, A. P. Lima & D. Cannatella (2014): Ciencias Exactas, Físicas y Naturales, 23: 303–316. Aposematism increases acoustic diversification and specia- Lynch, J. D., P. M. Ruíz-Carranza & M. C. Ardila-Robayo tion in poison frogs. – Proceedings of the Royal Society B, 218: (1997): Biogeographic patterns of Colombian frogs and toads. 20141761. – Revista Colombiana de la Academia de Ciencias Exactas, Schmidt, O. (1857): Diagnosen neuer Frösche des zoologischen Físicas y Naturales, 21: 237−248. Cabinets zu Krakau. – Sitzungsberichte der Kaiserlichen Aka- Martins, M. (1989): Nova especie de Colostethus da Amazonia demie der Wissenschaften, Mathematisch-Naturwissenschaft- Central (Amphibia: Dendrobatidae). – Revista Brasileira de liche Classe, 24: 10–15. Biologia, 49: 1009–1012. Steindachner, F. (1864): Batrachologische Mittheilungen. – Melin, D. E. (1941): Contributions to the knowledge of the Am- Verhandlungen des zoologisch-botanischen Vereins in Wien, phibia of South America. – Göteborgs Kungl. Vetenskaps-och 14: 239–288. Vitterhets-samhälles. Handlingar. Serien B, Matematiska och Sueur. J., T. Aubin & C. Simonis (2008): Seewave: a free modu- Naturvetenskapliga Skrifter,1 : 1–71. lar tool for sound analysis and synthesis. – Bioacoustics, 18: Noble, G. K. (1923): New batrachians from the Tropical Research 213–226. Station British Guiana. –Zoologica, 3: 289–299. Taylor, E. H. (1958): Additions to the known herpetological fau- Padial, J. M., S. Castroviejo-Fisher, J. Köhler, C. Vilà, J. C. na of Costa Rica with comments on other species. – University Chaparro & I. De la Riva (2009): Deciphering the products of Kansas Science Bulletin, 39: 3–40. of evolution at the species level: the need for an integrative tax- Toledo, L. F., I. A. Martins, D. P. Bruschi, M. A. Passos, C. Ale onomy. – Zoologica Scripta, 38: 431–447. xandre & C. F. B. Haddad (2015): The anuran calling reper- Pettit, B. A., G. R. Bourne & M. A. Bee (2012): Quantitative toire in the light of social context. – Acta Ethologica, 18: 87–99. acoustic analysis of the vocal repertoire of the Golden rocket Wells, K. D. (1980a): Behavioral ecology and social organization frog (Anomaloglossus beebei). – Journal of the Acoustical Soci- of a dendrobatid frog (Colostethus inguinalis). – Behavioral ety of America, 131: 4811–4820. Ecology and Sociobiology, 6: 199–209. Pröhl, H., S. Hagemann, J. Karsch & G. Höbel (2007): Geo- Wells, K. D. (1980b): Social behavior and communication of a graphic variation in male sexual signals in Strawberry poison dendrobatid frog (Colostethus trinitatis). – Herpetologica, 36: frogs (Dendrobates pumilio). – Ethology, 113: 825–837. 189–199. Quiguango-Ubillús, A. & L. A. Coloma (2008): Notes on be- Wells, K. D. (2007): The ecology and behavior of amphibians. – haviour, communication and reproduction in captive Hylo The University of Chicago Press, Chicago. xalus toachi (Anura: Dendrobatidae), an endangered Ecua- Werner, F. (1899): Ueber Reptilien und Batrachier aus Colum- dorian frog. – International Zoo Yearbook, 42: 78–89. bien und Trinidad. – Verhandlungen des zoologisch-bota- Rivero, J. A. (1963): Five new species of Atelopus from Colombia, nischen Vereins in Wien, 49: 470–484. with notes on other forms from Colombia and Ecuador. – Ca- Weygoldt, P. (1980): Zur Fortpflanzungsbiologie von Phyllo ribbean Journal of Science, 3: 103–124. bates femoralis (Boulenger) im Terrarium (Amphibia: Salien- Rivero, J. A. (1980 “1978”): Notas sobre los anfibios de Venezuela tia: Dendrobatidae). – Salamandra, 16: 215–226. III. Nuevos Colostethus de los Andes venezolanos. – Memoria. Zimmermann H. & E. Zimmermann (1985): Zur Fortpflan- Sociedad de Ciencias Naturales La Salle, 38: 95–111. zungsstrategie des Pfeilgiftfrosches Phyllobates terribilis My- Rojas-Runjaic, F. J. M., E. E. Infante-Rivero & C. L. Barrio- ers, Daly & Malkin, 1978. – Salamandra, 21: 281–297. Amorós (2011): A new frog of the genus Aromobates (Anura, Dendrobatidae) from Sierra de Perijá, Venezuela. – Zootaxa, Appendix 2919: 37–50. Examined specimens Rueda-Almonacid, J. V., A. A. Velásquez-Álvarez, P. A. Gal- vis-Peñuela & J. E. Gualdrón-Duarte (2008): Anfibios. – “Colostethus” ruthveni sensu stricto: COLOMBIA. Magda- pp. 169−192 in: Rodríguez-Mahecha, J. V., J. V. Rueda-Al- lena department: Santa Marta: Gaira river basin: corregimi- monacid & T. D. Gutiérrez-Hinojosa (eds): Guía ilustrada ento de Minca (CBUMAG:ANF:00253), Hacienda La Victo- de la fauna del Santuario de Vida Silvestre Los Besotes, Valle- ria (CBUMAG:ANF:00253, 527, 547–548, 551, 555–557, 560– dupar, Cesar, Colombia. – Serie de guías tropicales de campo 562, 567–568, 662); Toribio river basin: vereda Bella Vista Nº 7, Conservación Internacional-Colombia, Editorial Pana- (CBUMAG:ANF:00222); Los Rodríguez creek basin: vereda Las mericana, Formas e Impresos. Bogotá. Tinajas (CBUMAG:ANF:00752–757). “Colostethus” sp. ruthveni- Ruíz-Carranza, P. M., M. C. Ardila-Robayo & J. I. Hernán- like: COLOMBIA. Cesar department: Valledupar: corregimiento dez-Camacho (1994): Tres nuevas especies de Atelopus A. M. de Pueblo Bello (CBUMAG:ANF:891).“Colostethus cf. ruthveni: C. Duméril & Bibron 1841 (Amphibia: Bufonidae) de la Sierra COLOMBIA. La Guajira department. Dibulla: Reserva Wiman- Nevada de Santa Marta, Colombia. – Revista Colombiana de la gaga (CBUMAG:ANF:00939, 944, 948). Academia de Ciencias Exactas, Físicas y Naturales, 19: 153–163. Ruthven, A. G. (1922): The amphibians and reptiles of the Sierra Supplementary material Nevada de Santa Marta, Colombia. – Miscellaneous Publica- tions, Museum of Zoology, University of Michigan, 8: 1–69. Raw data and data of statistical analyses.